405
1 INTRODUCTION
Currently, in a number of cities of the Russian
Federation large projects are going on related to the
construction, modernization and expansion of
seaports.Thedevelopmentoftheshippingindustryis
oneofthekeydirectionsoftheeconomydevelopment
inthecountry.However,itentailsproblemssimilar
to
road traffic management. Since the maneuverability
of ships is significantly lower than that of the
automobiles, and weather conditions have a
considerably more significant impact, information
support systems for decision‐making in the
navigation safety domain are paid significant
attention. Currently, ships are equipped with
sophisticatedcontrolsystems,evaluating
thelocation
ofshipsaround,aswellasweatherconditions.Inthe
ports and places with particularly intensive marine
traffic, centralized navigation management systems
are used. However, integration and shared use of
information about marine traffic and upcoming
maneuvers of ships around to support decision‐
making when planning ship movement can
significantlyimprovethenavigationsafetyespecially
in places, which are not equipped with centralized
navigationmanagementsystems.
Besides, today it is a recognized fact that the
paradigm of information systems organization has
changed. This is caused by development of such
technologies as mobile Internet, knowledge
processing automation (cloud computing, crowd
sourcing,recommendersystems),andtheInternetof
Things (recently expanded to the notion of the
Internet of Everything – the Internet of physical
objects, IT services and humans). World‐leading
consulting companies estimate the pace of the
connection of new devices to the Internet as 5.5
milliondevicesperday(in
2020theInternetofThings
willinclude20.8billion devices [1]), andthemarket
volume for such systems by 2025 can exceed 20
trillionUSD[2].
These facts confirm the relevance of the new
research direction for tasks falling into the class of
System‐ of‐System engineering, since the amount
of
relationships types and relationships themselves
Recommender System for Navigation Safety: Requirements
and Methodology
N.Shilov
SPIIRAS,St.Petersburg,Russia
ABSTRACT: Low maneuverability of ships together with growing intensity of marine traffic result in new
challenges related to navigation safety. This paper reports a research aimed at design of methodology of
operationofrecommendersystemsfornavigationsafety.First,aspecificationofrequirementstosystemsofthe
considered
classhasbeencarriedout.Basedonthese,themajorprinciplesoffunctioningofsuchsystemshave
been defined. The principles were a basis for development of the mentioned above methodology, which is
basedontheusageofcontextpatternsandcharacterizedbythepresenceoffeedbacktoupdatethe
system’s
knowledgebase.
http://www.transnav.eu
the International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 14
Number 2
June 2020
DOI:10.12716/1001.14.02.18
406
greatly increases in such systems as well as
opportunitiesforpeopleandmachinestobenefitfrom
such relationships and available data. This requires
creation of new approaches, models, methods and
technologies for the entire diversity of socio‐cyber‐
physical systems (systems that tightly integrate
physicalworld,IT,andhumans)at
allstagesoftheir
lifecycle, taking into account their complexity and
dynamics caused by the variety of types of system
resourcesandchangingenvironmentalconditions.
Inthispaper,adecisionsupportsystemaimedat
support of navigation safety is proposed, which can
be considered as a typical example of a
distributed
socio‐cyber‐physical system operating in a dynamic
environment.Thepaperisstructuredasfollows.The
next section presents the state‐of‐the‐art in the
considered and adjacent areas. It is followed by
specification of requirements to the systems of the
consideredclass.Thentheprinciplesofoperationof
suchsystemsaredefinedbasedontherequirements.
The methodology built upon the principles is
presentedinthenextsection.Finally,themainresults
aresummarizedintheconclusion.
2 STATEOFTHEART
The intensive development of navigation, including
theNorthernSeaRoute,requiresconstantmonitoring
of the situation
from sides of both ground services
and ships themselves. As a result, it is necessary to
process large volumes of heterogeneous data.
Usually,monitoringsystemscollectinformationfrom
various sources and display it on digital maps to
facilitatethe workofdecisionmakers (inthispaper,
decisions stand for any actions
affecting the
development of the navigation situation), but in the
era of Big Data this is not enough. Intelligent data
analysis from heterogeneous sources is needed to
support decision making at a new level in order to
identifypotentiallydangeroussituations.
Information support of decision‐making for
navigationsafetycan
beimplementedinanumberof
ways. In particular, it could be a centralized system
managingmarinetrafficinaseaportandtheadjacent
waterarea[3],[4]orsystemswarningaboutpossible
emergency situations [5]–[7]. However, today there
arenointegrateddecisionsupportsystemsthatcould
takeintoaccountthe
currentsituationanditspossible
development not only for generation of warnings in
case of possible emergencies, but also for solving
navigationproblems.
Amongthesystems,whicharecloseenoughtothe
decision support systems in the area of navigation
safety, one can refer to the information systems
supportingthe
driverofmotorvehicles,whichhave
recently received a significant attention form the
research community. Some of them are focused on
improvingdrivingsafetythroughtrackingthestateof
the driver and his/her actions. Others are aimed at
supporting decision‐making related to the
convenience of the car usage (navigation, travel
planning,searchingforparkinglotsandsights,etc.).
One of the first companies to offer automatic
recognition of obstacles and collision avoidance was
Mobileye [8]. This company offers a number of
solutions aimed at interpreting the field of view
received by various cameras and sensors to prevent
collisions with animals,
pedestrians, vehicles and
other obstacles, as well as the recognition of road
signs and traffic lights. Today there are quite a
numberofsuchsolutions[9].Forexample,HARMAN
[10]offersvirtualandaugmentedrealitysystemsthat
allow the driver should “see” everything that
surroundsthecarinordertoincrease
drivingsafety,
aswellastoperformsuchmaneuversaslanechange
and parking. Considering the different needs of
different drivers, HARMAN solutions are aimed at
providinginformationinanintuitiveway,takinginto
accountthecurrentneedsofthedriver(forexample,
for local trips or for long journeys).
Unfortunately,
suchsystemscannotbeʺdirectlyʺusedfornavigation,
because they are very strongly tied to the specific
features of motor vehicles, however some of the
approaches and technologies can still be very useful
inthisarea.
Below, the existing methods and approaches
aimedatsolvingseparatetasks,whichareparts
ofthe
proposedresearchtopicareconsidered.
Since the system integrating information for
decision‐makingintheareaofnavigationsafetydeals
with dynamic environment, it has to be context‐
dependent,i.e.itsworkhastoconsiderprocessingthe
context of the current situation [11]. Context‐
dependence will not only
enable accounting for the
updatedinformationaboutthedynamicallychanging
situation, but also reducing the search space for
solutionsduetothedroppingoffinformationthatis
irrelevant to a particular situation. In this case the
context stands for all information describing the
current situation [12]. Context management
technology encompasses many
processes and
technologies that support information gathering,
management,andpublishinginanyform.Incontext‐
dependent decision support based on the usage of
ontology the information is usually presented in an
ontological form. Ontology management, in turn,
includes the processes of ontology creation through
integration of fragments of existing ontologies,
ontology maintenance, re‐using, adapting, and
evaluationoftheontologyforitscompliancewiththe
taskstobesolved.
Integrationofinformationfromvarioussourcesis
an integral part of distributed systems, that are the
classofsystemstheconsideredsystemfallsinto.The
problem of interoperability, i.e. the possibility of
varioussystemsanddatasourcestoworkjointlyare
usually divided into several levels [13]. The most
common model includes 3 levels: basic
interoperability, structural interoperability and
semanticinteroperability.
The level of basic (or functional,) interoperability
allows the data transmitted by one system to be
receivedbyanothersystem.This
level,however,does
notrequirethatthesystemonthereceivingsidecould
interpret the data. Basic compatibility can be
considered as the basis of the communication
pyramid, offering the most basic data exchange
services. To support the interoperability of the basic
layer, infrastructure solutions are used along with
various protocols
and data transfer standards [14]–
[16].
407
At the middle level, the structural (or syntactic)
interoperability determines the format of data
exchange that occurs between systems. This level is
mainly focused on defining message standards for
data transmission. In particular, structural
compatibility defines the syntax for data exchange
[17]–[19]. Since the content of a structured message
cannot
be standardized, a higher level of
interoperabilityisrequired.
Semantic interoperability is at the top of the
communicationpyramid.Thislevelisdefinedasthe
abilityofthesystemsnotonlytoshare,butalsotouse
the transmitted information. In this case, the
structuredmessagecontainsstandardizedcodeddata.
Thisallowsthereceivingsystemtointerpretthedata
[20].
In systems that include various independent
elements,semanticinteroperabilityiscrucialtobridge
the gap in terminology between them, as well as
between data sources. To solve this problem, it is
necessary to create a common dictionary that
provides accurate and
reliable communication
between the elements of the systems. And if the
interoperabilityatthefirsttwolevelstodayisusually
successfully solved due to the existence of different
standards, the problem of semantic interoperability
must be solved in each specific case because it is
impossible to define a single vocabulary
for all
existing information systems. As a rule, today,
ontology management technologies are usually used
to solve this problem [21], [22]. Ontologies are the
mostfrequentlyusedandtime‐testedtooltosupport
semantic interoperability, representing a set of
concepts of a certain problem area, as well as the
relationship between
them [23], [24]. Thus, in this
researchitwasdecidedtouseontologiesforsemantic
interoperability support within context‐driven
decision support systems in the field of navigation
safety.
To increase the efficiencyof processing incoming
information,itisproposedtousemethodsofcontext
management tied together with the
ontological
information representation. The context in this case
refers to a fragment of the ontology of the problem
area,includingelementsandrelationsbetweenthem
related to the current task, as well as instances of
ontology classes that have parameter values
correspondingtothecurrentsituation[25].
3 SPECIFICATIONOFREQUIREMENTS
TO
DECISIONSUPPORTSYSTEMSFOR
NAVIGATIONSAFETY
Based on the carried out analysis of the problem
domain, the following requirements to decision
support systems for navigation safety have been
specified:
information acquisition from distributed
heterogeneoussources,
flexibilityandscalabilitywithregardtothetypes
andquantityofinformation
sources,
responsiveness,
availability of means ensuring interoperability of
systemelements,
proactiveness,
applicationofrecommendersystemparadigm.
Below, each of the requirements is described in
detail.
3.1 Informationacquisitionfromdistributed
heterogeneoussources
Usually, navigation processes involve numerous
participants, which are often independent on each
other. These participants
include ships, services that
provideinformationabouttheweather,trafficonthe
waterways,portinfrastructure,etc.Thisleads to the
needtocollectandanalyzeinformationfromdifferent
owners, and, consequently, different formats and
semantics.
3.2 Flexibilityandscalabilitywithregardtothetypesand
quantityofinformationsources
This
requirement in some extent follows from the
previous one. Depending on how busy a particular
area is and if it has developed navigation
infrastructure, the number of information sources,
their availability, and the availability of certain type
of information may vary significantly. Decision
support systems working in the area of the
marine
traffic safety must be adapted to various operating
conditions without compromising the quality of
informationsupport.
3.3 Responsiveness
Oneofthecharacteristicsofthenavigationsituationis
itsvariability.Asaresult,decisionsupportsystemsin
the field of marine traffic safety must continuously
trackthechangingsituationand
respondtochanges,
whichcanleadtodangeroussituations.
3.4 Availabilityofmeansensuringinteroperabilityof
systemelements
Since, as already mentioned above, the operation of
decisionsupportsystemsintheareaofmarinetraffic
safetysupport,requiresprocessinginformationfrom
various independent sources, there is a problem of
supporting interoperability,
i.e. the possibility of
interaction of the system elements. In other words,
such a system should be able to extract information
fromvarioussources,processandintegrateit.
3.5 Proactiveness
Thedynamicdevelopmentofthesituationrelatedto
shiptraffic,aswellastheratherlongresponsetimeof
large
ships to control actions require in advance
implementation of actions aimed at preventing
dangeroussituations.Inthisregard,decisionsupport
systems in the field of marine safety should predict
thedevelopmentofthecurrentsituationandnotonly
provide information at the userʹs request, but also
408
behave proactively, i.e. they have to indicate the
possibility of a dangerous situation occurrence and
provide the information necessary for making a
decisiontopreventit.
3.6 Applicationofrecommendersystemparadigm
Asalreadymentioned,toensurethesafetyofmarine
traffic, it is necessary to analyze large amounts of
information and take into account many different
factors,anditisnotalwayspossibleforhumanstodo
this. Thus, it is proposed that the operation of
decisionsupportsystemsinthefieldofmarinesafety
should be based on the principles of recommender
systems, which do not only deliver
information
necessaryfordecision‐making,butalsoprovidefora
rankedlistofpreferreddecisions.Currently,thereare
a large number of models aimed at predicting the
development of the navigation situation, the use of
which for testing various solutions can significantly
improvethereliabilityoftheproposedsolutions,and,
consequently,
the level of marine safety. Existing
approaches to decision support in the area of
navigation safety are focused on providing
information about the current situation and the
applicationoftheparadigmofrecommendersystems
inthisareaisanewresult.
4 PRINCIPLESOFBUILDINGRECOMMENDER
SYSTEMSFORNAVIGATIONSAFETY
Based on the above requirements, the following
principles of building recommender systems for
navigationsafetyhavebeendefined:
serviceorientedarchitecture,
usage of ontologies for semantic interoperability
support,
applicationofcontextmanagementtechniques,
application of context templates to identify
potentiallydangeroussituationsandself‐learning.
The correspondence
between requirements and
principles fulfilling them is presented in Table 1.
Below,eachoftheprinciplesisdescribedindetail.
Table1.Correspondencebetweenrequirementsto
decision support systems for navigation safety and
principlesfulfillingthem.
4.1 ServiceOrientedArchitecture
Service‐oriented architecture allows not only using
external sources of
information with the ability to
connect and disconnect them, but also to scale and
adapt the functionality of the system without
affectingitsothercomponentsbyadding,eliminating
or duplicating the corresponding services. This
principleallowsthesystemtomeettherequirements
of1and2.
4.2 UsageofOntologiesfor
SemanticInteroperability
Support
Asaresultoftheanalysisofthestateoftheartinthis
area, it was concluded that ontologies are the most
promising apparatus for solving the problem of
semantic interoperability supporting when dealing
withheterogeneousindependentinformationsources.
However, the use of a single ontology
within the
system would be complicated due to the highly
dynamicinformationenvironment(astheshipmoves,
the sources of information constantly change). It is
proposed to solve this problem by fragmentation of
ontologies into specialized fragments (aspects).
Because of the aforementioned highly dynamic
information environment and the independence of
information
sources, it is either necessary to include
allpossibleterminologiesintotheontology,whichis
impossible,ortomatchtheexistingontologywiththe
ontology of the new information source each time.
However, since the manual ontology matching is a
very time and labor consuming process, and the
currently existing
methods of automatic ontology
matching are accurate only in narrow domains, the
fragmentation of the ontology of the considered
systemintospecializedaspectsisseenasalogicaland
effectivesolution that allows usage ofthe automatic
ontology matching methods. This principle makes it
possibleforthesystemtomeetthe
requirements1,3
and4.Theuseofmulti‐aspectontologiestoautomate
ontologymatchingwithintheinteroperabilitysupport
fordistributedsystemsisanewresult.
4.3 ApplicationofContextManagementTechniques
Themethodsofcontext management arefocusedon
determining which information is relevant for the
current task, and
which is not. Thus, these methods
allow,ontheonehand,increasingtheefficiencyofthe
system (requirement 3) by processing only the
information directly related to the current task, and
on the other hand, increasing the reliability of its
work by ensuring thatallinformation related to the
task is
taken into account. This principle allows the
systemtomeetrequirement3.
4.4 ApplicationofContextTemplatestoIdentify
PotentiallyDangerousSituationsandSelf‐Learning
Context templates [26] are partially instantiated
contexts of typical situations, for which it is known
whatactionsareneededtobetaken.Themainfeature
ofthefollowingthisprinciplewhencreatingasystem
isthepossibilityofaccumulatinghistoricaldataabout
the situations that have arisen together with the
actions taken, what allows further analysis. As a
result of the analysis, it can be determined which
characteristics are the most efficient for the given
navigation
situation and which of the actions taken
have led to its positive resolution. Thus, decision
supportsystemsinthefieldofmarinesafetygetthe
ability to self‐train, i.e. to accumulate the history of
typical situations (in the form of context templates)
that can lead to dangerous situations, as
well as
actionstobetakentopreventthedangeroussituation,
thatis,recommendations(requirements5and6).
409
Table1.Correspondencebetweenrequirementstodecisionsupportsystemsfornavigationsafetyandprinciplesfulfilling
them.
__________________________________________________________________________________________________
Principles
Requirements Information FlexibilityandResponsiveness Availabilityof Proactiveness Applicationof
acquisitionfromscalabilitywithmeansensuringrecommender
distributed regardtothetypesinteroperabilitysystem
heterogeneous andquantityofofsystemparadigm
sourcesinformationsourceselements
__________________________________________________________________________________________________
Serviceoriented ++
architecture
Usageof+++
ontologiesfor
semantic
interoperability
support
Applicationof+
context
management
techniques
Applicationof++
contexttemplates
toidentifypotentially
dangeroussituations
andself‐learning
__________________________________________________________________________________________________
5 METHODOLOGYOFOPERATIONOF
RECOMMENDERSYSTEMSFORNAVIGATION
SAFETY
This section proposes an original methodology of
operation of recommender systems for navigation
safety based on the principles defined above and
characterized by the presence of feedback to update
the system’s knowledge base. According to the
proposed methodology, the
operation of the
considered system aimed at identifying the general
preferencesofuser groups, takingintoaccounttheir
confidentiality,consistsofthefollowingsteps(Fig.1):
1 Search for available sources of information,
establishing communication channels with them
and ensuring semantic interoperability through
matching the relevant aspects of the system
ontology
and ontologies of the information
sources,whichinturnareformedonthebasisof
the analysis of the message structures from the
sources.
2 Collection of the information from sources,
translating it to the terminology of the system,
formation of the context of the current situation
andstoringit
inthecontextdatabase.
3 Comparisonofthecontextofthecurrentsituation
with known context patterns corresponding to
potentiallydangeroussituations.
4 Inthecaseofadetectedcorrespondencebetween
the current context and contexts stored in the
contexthistorytogetherwithactionsundertakenis
performed. Possible reasons for the
development
ofthesituationtoadangerousstateareidentified.
On the basis of the discovered patterns, new
patternsofcontextsandcorrespondingactionsare
created for both actions leading to dangerous
situationsandleadingto resolutionofpotentially
dangeroussituation.
5 Incaseofdeterminingadangeroussituation,the
user is offered a number of solutions
(recommendations) aimed at reducing the degree
of its danger, until a potentially safe situation is
reached.Thedecisionsmade(actionsundertaken)
arealsosavedinthesystem’sknowledgebase.
6 Ifthecurrentsituationissafe,thesystemcontinues
to monitor the decisions made
(if any) and the
currentsituation.
Figure1. Generic scheme of operation of Recommender
SystemsforNavigationSafety.
6 CONCLUSION
The paper proposes a novel approach to building
recommendersystemsfornavigationsafety.Basedon
the requirement specification major principles of
operation of such systems have been identified
including(i)informationacquisitionfromdistributed
heterogeneous sources, (ii) flexibility and scalability
withregardtothetypesandquantityofinformation
sources,(iii)responsiveness,(iv)availabilityofmeans
ensuring interoperability of system elements, (v)
proactiveness, (vi) application of recommender
system paradigm. Therequirements have been used
410
todefinethemethodologyoftheconsideredsystems
based on the usage of context patterns and
characterized by the presence of feedback to update
the system’s knowledge base. The future work is
aimed at prototyping of the approach in order to
investigate the most appropriate solutions for
particularaspectsof
theproposedmethodology.
ACKNOWLEDGEMENTS
Thereported studywas fundedby RFBRaccordingtothe
research project№18‐07‐01203 (state of the art analysis,
specification of requirements to decision support systems
fornavigationsafety,andthemethodologyofoperationof
theabovesystems)andbytheStateResearchno.0073‐2019‐
0005 (the recommender system
aspects and principles of
buildingrecommendersystemsfornavigationsafety).
REFERENCES
[1]R.vanderMeulen,“GartnerSays6.4BillionConnected
‘Things’ Will Be in Use in 2016, Up 30 Percent From
2015,” 2015. [Online]. Available:
https://www.gartner.com/en/newsroom/press‐
releases/2015‐11‐10‐gartner‐says‐6‐billion‐connected‐
things‐will‐be‐in‐use‐in‐2016‐up‐30‐ percent‐from‐2015.
[Accessed:26‐Mar‐2019].
[2]
J.Manyika,M.Chui,J.Bughin,R.Dobbs,P.Bisson,and
A.Marrs, “Disruptivetechnologies: Advancesthat will
transformlife,business,andtheglobaleconomy,”2013.
[3]S.Y.Huang,W.J.Hsu,H.Fang,andT.Song,“MTSS‐‐
A Marine Traffic Simulation System and Scenario
Studies for a Major
Hub Port,” ACM Trans. Model.
Comput.Simul.,vol.27,no.1,pp.1–26,Aug.2016.
[4]B.Dragović, E.Tzannatos,andN. K.Park,“Simulation
modelling in ports and container terminals: literature
overviewandanalysisbyresearchfield,applicationarea
andtool,”Flex.Serv.Manuf.J.,vol.29,no.
1,pp.4–34,
Mar.2017.
[5]S.‐L. Kao, K.‐Y. Chang, and T.‐W. Hsu, “Fuzzy
groundingalert systemforvessel trafficservice via3D
marineGIS,”J.Mar.Sci.Technol.,vol.25,no.2,pp.186–
195,2017.
[6]E. Krishnamoorthy, S. Manikandan, M. M. Shalik, and
M. Kumarasamy,
“Border alert system and emergency
contactforfishermanusingGPS,”Int.J.Res.Eng.,vol.4,
no.3,pp.66–68,2017.
[7]E.Pratiwi,K.B.Artana,andA.A.B.Dinariyana,“Fuzzy
InferenceSystemforDeterminingCollisionRiskofShip
in Madura Strait Using Automatic Identification
System,”Int.J.
Mar.Environ.Sci.,vol.11,no.2,pp.401–
405,2017.
[8]“Mobileye,” 2019. [Online]. Available:
https://www.mobileye.com/.[Accessed:26‐Mar‐2019].
[9]A. Taramov and N. Shilov, “A systematic review of
proactive driver support systems and underlying
technologies,” in 2017 20th Conference of Open
InnovationsAssociation(FRUCT),2017,pp.448–459.
[10]Harman,“AdvancedDriverAssistanceSystems,”2019.
[Online]. Available: https://car.harman.com/solutions/
advanced‐driver‐assistance‐systems. [Accessed:26‐
Mar‐2019].
[11]A.Smirnov,T.Levashova,N.Shilov,andA.Kashevnik,
“Hybridtechnologyforself‐organizationofresourcesof
pervasive environment for operational decision
support,” Int. J. Artif. Intell. Tools, vol. 19, no. 2,
pp.
211–229,2010.
[12]A.K.Dey,G.D.Abowd,andD.Salber,“AConceptual
Framework and a Toolkit for Supporting the Rapid
Prototyping of Context‐Aware Applications,” Human–
Computer Interact., vol. 16, no. 2–4, pp. 97–166, Dec.
2001.
[13]H. Dibowski, “Semantic interoperability evaluation
modelfordevicesinautomationsystems,”
in201722nd
IEEE International Conference on Emerging
TechnologiesandFactoryAutomation(ETFA),2017,pp.
1–6.
[14]P.‐P. J. Beaujean, E. A. Carlson, J. Spruance, and D.
Kriel, “HERMES‐A high‐speed acoustic modem for
real‐time transmission of uncompressed image and
status transmission in port environment and very
shallowwater,”inOCEANS2008,2008,pp.1–9.
[15]D. Locke, “MQ Telemetry Transport (MQTT) V3.1
ProtocolSpecification,”2010.
[16]J.Rodríguez‐Molina,B.Martínez,S.Bilbao,andT.
Martín‐Wanton, “Maritime Data Transfer Protocol
(MDTP):AProposalforaDataTransmissionProtocolin
Resource‐Constrained Underwater Environments
InvolvingCyber‐Physical
Systems,”Sensors,vol.17,no.
6,p.1330,Jun.2017.
[17]W3C,“SOAPVersion1.2Part1:MessagingFramework
(Second Edition),” W3C Recommendation, 2007.
[Online]. Available: https://www.w3.org/TR/soap12/.
[Accessed:26‐Mar‐2019].
[18]W3C,“ExtensibleMarkupLanguage(XML)1.0(Fifth
Edition),” W3CRecommendation, 2008. [Online].
Available: https://www.w3.org/TR/xml/. [Accessed: 26‐
Mar‐2019].
[19]ECMA‐International, “The JSON Data Interchange
Syntax,” 2017. [Online]. Available: http://www.ecma‐
international.org/publications/files/ECMA‐ST/ECMA‐
404.pdf.[Accessed:26‐Mar‐2019].
[20]P.G.Larsen,J.Fitzgerald,J.Woodcock,R.Nilsson,C.
Gamble, and S. Foster, “Towards Semantically
IntegratedModelsandToolsforCyber‐PhysicalSystems
Design,” in ISoLA2016: Leveraging Applications of
Formal Methods, Verification and Validation:
Discussion,Dissemination, Applications. LectureNotes
inComputerScience,Springer,2016,pp.171–186.
[21]G.DeGiacomo,D.Lembo,M.Lenzerini,A.Poggi,and
R. Rosati, “Using Ontologies for Semantic Data
Integration,” in A Comprehensive Guide Through the
Italian Database Research Over the Last 25 Years.
StudiesinBigData,Springer,2018,pp.187–202.
[22]Z. Wang, N. Chen, W. Zhang, F. Pu, and C. Du,
“Semantic integration of wireless sensor networks into
open geospatial consortium sensor observation service
toaccessandshareenvironmentalmonitoringsystems,”
IETSoftw.,vol.10,no.2,pp.45–53,Apr.2016.
[23]
F. Carrez, T. Elsaleh, D. Gomez, L. Sanchez, J. Lanza,
and P.Grace, “A ReferenceArchitecturefor federating
IoT infrastructures supporting semantic
interoperability,” in 2017 European Conference on
NetworksandCommunications(EuCNC),2017,pp.1–6.
[24]A.Gyrard, S.K. Datta,andC.Bonnet, “Asurveyand
analysis of ontology‐
based software tools for semantic
interoperability in IoT and WoT landscapes,” in 2018
IEEE4thWorldForumonInternetofThings(WF‐IoT),
2018,pp.86–91.
[25]K. Sandkuhl and A. Smirnov, “Context‐oriented
Knowledge Management in Production Networks,”
Appl.Comput.Syst.,vol.23,no.2,pp.81–89,Dec.2018.
[26]
A. Smirnov, K. Sandkuhl, N. Shilov, and N. Teslya,
“Service Self‐Contextualization in Cyber‐Physical
Systems based on Context Modeling and Context
Variation,” in Joint Proceedings of the BIR 2018 Short
Papers,WorkshopsandDoctoralConsortiumco‐located
with 17th International Conference Perspectives in
BusinessInformaticsResearch(BIR2018),2018,
pp.94–
105.
